Skip to main content

John MacIntyre

University of Sunderland, International Management, Pro Vice Chancellor

University of Sunderland, Applied Sciences, Faculty Member

Followers

73

Following

4

Co-authors

4

Public Views

Senior academic executive, former Dean of Faculty of Applied Sciences and now Pro Vice Chancellor (International) responsible for a large international operation covering student recruitment, academic partnerships, and overseas campus operations.Research background in applied artificial intelligence, as well as leadership and business change management. Editor-in-Chief of Neural Computing

less

Christos Emmanouilidis

University of Groningen

Kenneth McGarry

University of Sunderland

Tallinn University of Technology

Nicolò Bachschmid

Penn State University Berks

University of Oulu

Interests

Uploads

Papers by John MacIntyre

Rotating Machinery Fault Diagnosis Based on Neurofuzzy Computing

This paper describes the application of neurofuzzy computing techniques for fault diagnosis in ro... more This paper describes the application of neurofuzzy computing techniques for fault diagnosis in rotating machinery. It forms part of the work currently undertaken by the VISION project, a European research project, which aims at developing an intelligent diagnostic system, based on the fusion of various simulation and computational intelligence techniques. Vibration data is often or can easily become available from rotating machinery, but in many cases there is little provision for taking full advantage of its potential for improving reliability in machine condition diagnostics. There is also ample domain knowledge about machinery malfunctioning, but fault diagnosis is too complex to be reliably provided by simple rules. The main difficulties are due to measurement noise, uncertainty about the relationship between the measured quantities and the machine condition, the dumping of the signal over the transmission path and differences in machine configuration. There is a clear need for ...

Neurofuzzy Computing Aided Machine Fault Diagnosis

Machinery malfunction problems are often sources of increased maintenance costs and disturbances ... more Machinery malfunction problems are often sources of increased maintenance costs and disturbances in production activity across industry. Reliable diagnostic methodologies are needed to enable cost effective condition based maintenance. Among the methods that can be employed for fault diagnosis, neurofuzzy techniques offer an appealing problem solving framework. This is due to their capability for simultaneously handling numerical and linguistic information. This paper demonstrates the use of neurofuzzy techniques for machinery fault diagnosis, based on measuring vibration. Data driven diagnostic model building is facilitated by the availability of rich vibration data sources, including simulation and real data gathered from purpose-built test rigs and machinery equipment at industrial sites. This work is carried out as part of the VISION project, a European research project, which is developing a generic machinery fault diagnostic system, based on a fusion of computational intellige...

Adaptive Fusion Devices for Condition Monitoring: An Overview of the NEURAL-MAINE Project

Key Engineering Materials

Knowledge extraction and insertion from radial basis function networks

IEE Colloquium on Applied Statistical Pattern Recognition

ABSTRACT Neural networks provide excellent solutions for pattern recognition and classification p... more ABSTRACT Neural networks provide excellent solutions for pattern recognition and classification problems. Unfortunately, in the case of distributed neural networks such as the multilayer perceptron it is difficult to comprehend the learned internal mappings. This makes any form of explanation facility such as that possessed by expert systems impractical. However, in the case of localist neural representations the situation is more transparent to examination. This paper examines the quality and comprehensibility of rules extracted from localist neural networks, specifically the radial basis function network. The rules are analysed in order to gain knowledge and insight into the data. We also investigate the benefits of inserting prior domain knowledge into a radial basis function network

Stepwise Linear Regression for Dimensionality Reduction in Neural Network Modelling

This work considers the applicability of applying the derivatives of stepwise linear regression m... more This work considers the applicability of applying the derivatives of stepwise linear regression modelling (specifically the p-values which indicate the importance of a variable to the modelling process) as a feature extraction technique. We utilise it in conjunction with several data sets of varying levels of complexity, and compare our results to other dimensionality reduction techniques such as genetic algorithms,

Methods for integrating memory into neural networks in condition monitoring

A criticism of neural network architectures is their susceptibility to "catastrophic interference... more A criticism of neural network architectures is their susceptibility to "catastrophic interference" the ability to forget previously learned data when presented with new patterns. To avoid this, neural network architectures have been developed which specifically provide the network with a memory, either through the use of a context unit, which can store patterns for later recall, or which combine high-levels of recurrency coupled with some form of backpropagation. We have evaluated two architectures which utilise these concepts, namely, Hopfield and Elman networks, respectively and compared their performance to self-organising feature maps using time-smoothed moving average data and Time delayed neural networks. Our results indicate clear improvements in performance for networks incorporating memory into their structure. However the degree of improvement depends largely upon the architecture used, and the provision of a context layer for the storage and recall of patterns.

An Investigation into the perceived barriers to total productive maintenance in small medium enterprises

Knowledge extraction from local function networks

Proceedings of the 17th International Joint Conference on Artificial Intelligence Volume 2, Jul 12, 2001

Extracting rules from RBFs is not a trivial task because of nonlinear functions or high input dim... more Extracting rules from RBFs is not a trivial task because of nonlinear functions or high input dimensionality. In such cases, some of the hidden units of the RBF network have a tendency to be "shared" across several output classes or even may not contribute to any output class. To address this we have developed an algorithm called LREX (for Local Rule EXtraction) which tackles these issues by extracting rules at two levels: hREX extracts rules by examining the hidden unit to class assignments while mREX extracts rules based on the input space to output space mappings. The rules extracted by our algorithm are compared and contrasted against a competing local rule extraction system. The central claim of this paper is that local function networks such as radial basis function (RBF) networks have a suitable architecture based on Gaussian functions that is amenable to rule extraction.

Modern Maintenance Practices: The barriers to implementation within small/medium enterprises

The fusion of neuro-symbolic knowledge within an adaptive hybrid system

Data Mining and User Profiling for An E-Commerce System. Proceedings of the 1st International Conference on Fuzzy Systems and Knowledge Discovery

Fuzzy Systems and Knowledge Discovery, 2002

<title>Adaptive local fusion systems for novelty detection and diagnostics in condition monitoring</title>

Sensor Fusion: Architectures, Algorithms, and Applications II, 1998

ABSTRACT This paper describes the application of Kohonen Self Organizing Maps in a dynamic machin... more ABSTRACT This paper describes the application of Kohonen Self Organizing Maps in a dynamic machine condition monitoring application which learns fault conditions over time. The authors describe the implementation of a novelty detection and adaptive diagnostic system which forms a modular component of a larger on-line condition monitoring system.

Adaptive local fusion systems for novelty detection and diagnostics in condition monitoring

Storage and Retrieval for Image and Video Databases, 1998

This paper describes the application of Kohonen Self Organizing Maps in a dynamic machine conditi... more This paper describes the application of Kohonen Self Organizing Maps in a dynamic machine condition monitoring application which learns fault conditions over time. The authors describe the implementation of a novelty detection and adaptive diagnostic system which forms a modular component of a larger on-line condition monitoring system.

Hybrid diagnostic system based upon simulation and artificial intelligence

The authors have developed an intelligent, adaptive diagnostic hybrid system for the interpretati... more The authors have developed an intelligent, adaptive diagnostic hybrid system for the interpretation of vibration data for the condition monitoring of rotating machinery. The main objective was to produce a diagnostic system capable of ensuring a high level of machine reliability across a wide variety of potential plant configurations. The diagnostic hybrid system uses a variety of intelligent techniques such as neural networks, neurofuzzy logic and rule based reasoning to enable accurate and reliable predictions of machine health to be made. Such intelligent techniques have been successfully used in other competing systems, however existing systems are generally dedicated to specific items of plant and are not adaptive. In order to achieve a generic diagnostic system it was necessary to develop a simulation model capable of representing the underlying mathematical features of a basic plant configuration. The use of data from a specially constructed test rig and real-world data gathered from user sites was used to refine the initial model. As the accuracy and confidence of the simulation model increased more sophisticated configurations could be generated thus providing the data necessary for training the neural networks and neurofuzzy systems.

Measuring the alignment of Maintenance andManufacturing Strategies – The developmentof a new model and diagnostic tool

Purpose-To outline the development of a new conceptual model and diagnostic tool which assesses t... more Purpose-To outline the development of a new conceptual model and diagnostic tool which assesses the strategic processes and links involved between maintenance and manufacturing strategies and the effect on manufacturing performance. Methodology-The new model was developed in two stages. Firstly a conceptual framework was produced and then a holistic diagram. From this theoretical model a diagnostic tool was generated. This tool was then verified and tested during an empirical research phase which produced four extensive and detailed case studies. A brief overview of one of the case studies is provided in this article. Findings-It was found that the conceptual model provided an accurate representation of the strategic processes and links which should exist in a manufacturing plant. Moreover the diagnostic tool proved to be a valid and reliable test instrument to measure and display this information. Practical implications-The diagnostic tool provides a snapshot of the status of manufacturing and maintenance strategies which then can be used to benchmark improvements over time. If required the individual questions from the questionnaire can be "reverse engineered" to provide detailed information for senior managers to carry out corrective action. Originality/value-This theoretical and empirical research closes a gap in the literature relating to the linkage between maintenance and manufacturing strategies. It does so by providing a unique and holistic model showing the strategic processes and links which should exist within a manufacturing plant. Moreover the diagnostic tool produced from the model is a convenient audit facility which enables companies to move toward functional coherence.

Knowledge Transfer between Neural Networks

Modelling Dynamics Control Systems to Improve Crash Characteristics and Occupant Behaviour

Improved Classification for a Data Fusing Kohonen Self Organizing Map Using A Dynamic Thresholding Technique

International Joint Conference on Artificial Intelligence, 1999

The use of linear data fusion is a fast develop ing area in the field of military information an... more The use of linear data fusion is a fast develop ing area in the field of military information and combat systems. However, the use of data fusion in conventional application areas is not as wide spread. To date linear data fusion has been used only in applications in which substantial knowledge of both the problem domain and the sensor devices

Knowledge extraction and insertion from radial basis function networks

Neural networks provide excellent solutions for pattern recognition and classification problems. ... more Neural networks provide excellent solutions for pattern recognition and classification problems. Unfortunately, in the case of distributed neural networks such as the multilayer perceptron it is difficult to comprehend the learned internal mappings. This makes any form of explanation facility such as that possessed by expert systems impractical. However, in the case of localist neural representations the situation is more transparent to examination. This paper examines the quality and comprehensibility of rules extracted from localist neural networks, specifically the radial basis function network. The rules are analysed in order to gain knowledge and insight into the data. We also investigate the benefits of inserting prior domain knowledge into a radial basis function network

Two Biologically Plausible Network Architectures for the Avoidance of Catastrophic Interference

This paper considers the current state of the art regarding the avoidance of catastrophic interfe... more This paper considers the current state of the art regarding the avoidance of catastrophic interference, a phenomena experienced by neural network architectures which share weights. The paper compares and contrasts the authors Biologically plausible neural network architecture (termed Biologically plausible model, or BPM) to the pseudo recurrent neural networks developed by Robert French using two real world data sets which have been used in previous studies of catastrophic interference. The results show that although the pseudo recurrent techniques do offer significant advantages in recall performance, the authors model does

Rotating Machinery Fault Diagnosis Based on Neurofuzzy Computing

This paper describes the application of neurofuzzy computing techniques for fault diagnosis in ro... more This paper describes the application of neurofuzzy computing techniques for fault diagnosis in rotating machinery. It forms part of the work currently undertaken by the VISION project, a European research project, which aims at developing an intelligent diagnostic system, based on the fusion of various simulation and computational intelligence techniques. Vibration data is often or can easily become available from rotating machinery, but in many cases there is little provision for taking full advantage of its potential for improving reliability in machine condition diagnostics. There is also ample domain knowledge about machinery malfunctioning, but fault diagnosis is too complex to be reliably provided by simple rules. The main difficulties are due to measurement noise, uncertainty about the relationship between the measured quantities and the machine condition, the dumping of the signal over the transmission path and differences in machine configuration. There is a clear need for ...

Neurofuzzy Computing Aided Machine Fault Diagnosis

Machinery malfunction problems are often sources of increased maintenance costs and disturbances ... more Machinery malfunction problems are often sources of increased maintenance costs and disturbances in production activity across industry. Reliable diagnostic methodologies are needed to enable cost effective condition based maintenance. Among the methods that can be employed for fault diagnosis, neurofuzzy techniques offer an appealing problem solving framework. This is due to their capability for simultaneously handling numerical and linguistic information. This paper demonstrates the use of neurofuzzy techniques for machinery fault diagnosis, based on measuring vibration. Data driven diagnostic model building is facilitated by the availability of rich vibration data sources, including simulation and real data gathered from purpose-built test rigs and machinery equipment at industrial sites. This work is carried out as part of the VISION project, a European research project, which is developing a generic machinery fault diagnostic system, based on a fusion of computational intellige...

Adaptive Fusion Devices for Condition Monitoring: An Overview of the NEURAL-MAINE Project

Key Engineering Materials

Knowledge extraction and insertion from radial basis function networks

IEE Colloquium on Applied Statistical Pattern Recognition

ABSTRACT Neural networks provide excellent solutions for pattern recognition and classification p... more ABSTRACT Neural networks provide excellent solutions for pattern recognition and classification problems. Unfortunately, in the case of distributed neural networks such as the multilayer perceptron it is difficult to comprehend the learned internal mappings. This makes any form of explanation facility such as that possessed by expert systems impractical. However, in the case of localist neural representations the situation is more transparent to examination. This paper examines the quality and comprehensibility of rules extracted from localist neural networks, specifically the radial basis function network. The rules are analysed in order to gain knowledge and insight into the data. We also investigate the benefits of inserting prior domain knowledge into a radial basis function network

Stepwise Linear Regression for Dimensionality Reduction in Neural Network Modelling

This work considers the applicability of applying the derivatives of stepwise linear regression m... more This work considers the applicability of applying the derivatives of stepwise linear regression modelling (specifically the p-values which indicate the importance of a variable to the modelling process) as a feature extraction technique. We utilise it in conjunction with several data sets of varying levels of complexity, and compare our results to other dimensionality reduction techniques such as genetic algorithms,

Methods for integrating memory into neural networks in condition monitoring

A criticism of neural network architectures is their susceptibility to "catastrophic interference... more A criticism of neural network architectures is their susceptibility to "catastrophic interference" the ability to forget previously learned data when presented with new patterns. To avoid this, neural network architectures have been developed which specifically provide the network with a memory, either through the use of a context unit, which can store patterns for later recall, or which combine high-levels of recurrency coupled with some form of backpropagation. We have evaluated two architectures which utilise these concepts, namely, Hopfield and Elman networks, respectively and compared their performance to self-organising feature maps using time-smoothed moving average data and Time delayed neural networks. Our results indicate clear improvements in performance for networks incorporating memory into their structure. However the degree of improvement depends largely upon the architecture used, and the provision of a context layer for the storage and recall of patterns.

An Investigation into the perceived barriers to total productive maintenance in small medium enterprises

Knowledge extraction from local function networks

Proceedings of the 17th International Joint Conference on Artificial Intelligence Volume 2, Jul 12, 2001

Extracting rules from RBFs is not a trivial task because of nonlinear functions or high input dim... more Extracting rules from RBFs is not a trivial task because of nonlinear functions or high input dimensionality. In such cases, some of the hidden units of the RBF network have a tendency to be "shared" across several output classes or even may not contribute to any output class. To address this we have developed an algorithm called LREX (for Local Rule EXtraction) which tackles these issues by extracting rules at two levels: hREX extracts rules by examining the hidden unit to class assignments while mREX extracts rules based on the input space to output space mappings. The rules extracted by our algorithm are compared and contrasted against a competing local rule extraction system. The central claim of this paper is that local function networks such as radial basis function (RBF) networks have a suitable architecture based on Gaussian functions that is amenable to rule extraction.

Modern Maintenance Practices: The barriers to implementation within small/medium enterprises

The fusion of neuro-symbolic knowledge within an adaptive hybrid system

Data Mining and User Profiling for An E-Commerce System. Proceedings of the 1st International Conference on Fuzzy Systems and Knowledge Discovery

Fuzzy Systems and Knowledge Discovery, 2002

<title>Adaptive local fusion systems for novelty detection and diagnostics in condition monitoring</title>

Sensor Fusion: Architectures, Algorithms, and Applications II, 1998

ABSTRACT This paper describes the application of Kohonen Self Organizing Maps in a dynamic machin... more ABSTRACT This paper describes the application of Kohonen Self Organizing Maps in a dynamic machine condition monitoring application which learns fault conditions over time. The authors describe the implementation of a novelty detection and adaptive diagnostic system which forms a modular component of a larger on-line condition monitoring system.

Adaptive local fusion systems for novelty detection and diagnostics in condition monitoring

Storage and Retrieval for Image and Video Databases, 1998

This paper describes the application of Kohonen Self Organizing Maps in a dynamic machine conditi... more This paper describes the application of Kohonen Self Organizing Maps in a dynamic machine condition monitoring application which learns fault conditions over time. The authors describe the implementation of a novelty detection and adaptive diagnostic system which forms a modular component of a larger on-line condition monitoring system.

Hybrid diagnostic system based upon simulation and artificial intelligence

The authors have developed an intelligent, adaptive diagnostic hybrid system for the interpretati... more The authors have developed an intelligent, adaptive diagnostic hybrid system for the interpretation of vibration data for the condition monitoring of rotating machinery. The main objective was to produce a diagnostic system capable of ensuring a high level of machine reliability across a wide variety of potential plant configurations. The diagnostic hybrid system uses a variety of intelligent techniques such as neural networks, neurofuzzy logic and rule based reasoning to enable accurate and reliable predictions of machine health to be made. Such intelligent techniques have been successfully used in other competing systems, however existing systems are generally dedicated to specific items of plant and are not adaptive. In order to achieve a generic diagnostic system it was necessary to develop a simulation model capable of representing the underlying mathematical features of a basic plant configuration. The use of data from a specially constructed test rig and real-world data gathered from user sites was used to refine the initial model. As the accuracy and confidence of the simulation model increased more sophisticated configurations could be generated thus providing the data necessary for training the neural networks and neurofuzzy systems.

Measuring the alignment of Maintenance andManufacturing Strategies – The developmentof a new model and diagnostic tool

Purpose-To outline the development of a new conceptual model and diagnostic tool which assesses t... more Purpose-To outline the development of a new conceptual model and diagnostic tool which assesses the strategic processes and links involved between maintenance and manufacturing strategies and the effect on manufacturing performance. Methodology-The new model was developed in two stages. Firstly a conceptual framework was produced and then a holistic diagram. From this theoretical model a diagnostic tool was generated. This tool was then verified and tested during an empirical research phase which produced four extensive and detailed case studies. A brief overview of one of the case studies is provided in this article. Findings-It was found that the conceptual model provided an accurate representation of the strategic processes and links which should exist in a manufacturing plant. Moreover the diagnostic tool proved to be a valid and reliable test instrument to measure and display this information. Practical implications-The diagnostic tool provides a snapshot of the status of manufacturing and maintenance strategies which then can be used to benchmark improvements over time. If required the individual questions from the questionnaire can be "reverse engineered" to provide detailed information for senior managers to carry out corrective action. Originality/value-This theoretical and empirical research closes a gap in the literature relating to the linkage between maintenance and manufacturing strategies. It does so by providing a unique and holistic model showing the strategic processes and links which should exist within a manufacturing plant. Moreover the diagnostic tool produced from the model is a convenient audit facility which enables companies to move toward functional coherence.

Knowledge Transfer between Neural Networks

Modelling Dynamics Control Systems to Improve Crash Characteristics and Occupant Behaviour

Improved Classification for a Data Fusing Kohonen Self Organizing Map Using A Dynamic Thresholding Technique

International Joint Conference on Artificial Intelligence, 1999

The use of linear data fusion is a fast develop ing area in the field of military information an... more The use of linear data fusion is a fast develop ing area in the field of military information and combat systems. However, the use of data fusion in conventional application areas is not as wide spread. To date linear data fusion has been used only in applications in which substantial knowledge of both the problem domain and the sensor devices

Knowledge extraction and insertion from radial basis function networks

Neural networks provide excellent solutions for pattern recognition and classification problems. ... more Neural networks provide excellent solutions for pattern recognition and classification problems. Unfortunately, in the case of distributed neural networks such as the multilayer perceptron it is difficult to comprehend the learned internal mappings. This makes any form of explanation facility such as that possessed by expert systems impractical. However, in the case of localist neural representations the situation is more transparent to examination. This paper examines the quality and comprehensibility of rules extracted from localist neural networks, specifically the radial basis function network. The rules are analysed in order to gain knowledge and insight into the data. We also investigate the benefits of inserting prior domain knowledge into a radial basis function network

Two Biologically Plausible Network Architectures for the Avoidance of Catastrophic Interference

This paper considers the current state of the art regarding the avoidance of catastrophic interfe... more This paper considers the current state of the art regarding the avoidance of catastrophic interference, a phenomena experienced by neural network architectures which share weights. The paper compares and contrasts the authors Biologically plausible neural network architecture (termed Biologically plausible model, or BPM) to the pseudo recurrent neural networks developed by Robert French using two real world data sets which have been used in previous studies of catastrophic interference. The results show that although the pseudo recurrent techniques do offer significant advantages in recall performance, the authors model does